for the past few months I’ve been trying to troubleshoot an issue were the core restarts for no reason or one of the clients (roon bridge) stops playing music.
I have realized that it happens when the computers (core and bridge) synchronize the clock via NTP. It seems that if the core or bridge are playing music and the clock adjusts due to time drift the core/bridge crashes and restarts.
The core is running on Ubuntu 18.04 the bridge is running on DietPi (latest).
The roon bridge log shows a bunch of binary data in the log after this happens, the roon core log shows a null pointer exception and no stack trace.
You should be able to reproduce this by just setting the clock manually while playing music.
Please do something to fix this, or at least add an option to restart music that was playing when this happens.
Thanks for contacting support and letting us know more about this issue that you are facing. I have started a case for you with our QA team who will investigate this issue, I will be sure to let you know what they say once they complete their report and pass it back to me. I appreciate your patience in the meantime.
I have just received the report back from QA and they have unfortunately not been able to reproduce the issue that you are facing. They have asked if you can clarify the steps you took to have this issue occur and be reproducible, I have the following so far, please correct me if I am wrong:
Have the Roon Core and Bridge Open
Have music playing to the Roon Bridge (DietPi)
Manually sync NTP time on the Ubuntu Core
This is the step when the song theoretically cuts out
Are you able to manually reproduce this issue consistently when you sync your NTP time with the Core? I have gone ahead and enabled diagnostics mode for your account and what this action will do is next time the Core is active, a set of logs will automatically be generated and uploaded to our servers for further analysis.
I kindly ask for you to reproduce this issue once more and note the exact local time in your country (ex 1:00PM) that you notice the music cut out. Once we have the timestamps, we can cross-reference the logs to see if there is something going on in the background which should not be happening. If you have any more information as to expand on exactly what the steps you used are please let us know.
Sure, I’m traveling this week but I will try to reproduce this over the weekend.
Just to clarify, the NTP sync itself does not seem to the problem, I believe the issue occurs when the NTP sync actually has to adjust the clock to bring it in line with the remote servers.
Thanks for confirming when the issue occurs. If you are able to get back into this state by manually changing the clock for either the Core or the Bridge please provide reproduction steps when you are back from your travels and we will take a look at the issue and attempt to reproduce in the lab.
Sorry for the late reply, I’ve been testing this more and it seems that the clock change sometimes causes a network restart, not always but sometimes and of course that kills the audio stream. I have not been able to reproduce the core/bridge crash.
So I’ve implemented a controller to monitor and restart playback if it detects network loss or a zone/core loss.
You can close the ticket and revert logging to normal level.